arXiv subjects

Carlos Gershenson

Publications and source records attributed to Carlos Gershenson.

At least 73 records · Page 4Linked to original sources

arXiv2013

Measuring the Complexity of Ultra-Large-Scale Adaptive Systems

Ultra-large scale (ULS) systems are becoming pervasive. They are inherently complex, which makes their design and control a challenge for traditional methods. Here we propose the design and analysis of ULS systems using measures of complexity, emergence, self-organization, and homeostasis based on information theory. These measures allow the evaluation of ULS systems and thus can be used to guide their design. We evaluate the proposal with a ULS computing system provided with adaptation mechanisms. We show the evolution of the system with stable and also changing workload, using different fitness functions. When the adaptive plan forces the system to converge to a predefined performance level, the nodes may result in highly unstable configurations, that correspond to a high variance in time of the measured complexity. Conversely, if the adaptive plan is less "aggressive", the system may be more stable, but the optimal performance may not be achieved.

cs.NE↗

arXiv2013

Self-Organization Promotes the Evolution of Cooperation with Cultural Propagation

In this paper three computational models for the study of the evolution of cooperation under cultural propagation are studied: Kin Selection, Direct Reciprocity and Indirect Reciprocity. Two analyzes are reported, one comparing their behavior between them and a second one identifying the impact that different parameters have in the model dynamics. The results of these analyzes illustrate how game transitions may occur depending of some parameters within the models and also explain how agents adapt to these transitions by individually choosing their attachment to a cooperative attitude. These parameters regulate how cooperation can self-organize under different circumstances. The emergence of the evolution of cooperation as a result of the agent's adapting processes is also discussed.

cs.GT↗

arXiv2013

Living is information processing: from molecules to global systems

We extend the concept that life is an informational phenomenon, at every level of organisation, from molecules to the global ecological system. According to this thesis: (a) living is information processing, in which memory is maintained by both molecular states and ecological states as well as the more obvious nucleic acid coding; (b) this information processing has one overall function - to perpetuate itself; and (c) the processing method is filtration (cognition) of, and synthesis of, information at lower levels to appear at higher levels in complex systems (emergence). We show how information patterns, are united by the creation of mutual context, generating persistent consequences, to result in `functional information'. This constructive process forms arbitrarily large complexes of information, the combined effects of which include the functions of life. Molecules and simple organisms have already been measured in terms of functional information content; we show how quantification may be extended to each level of organisation up to the ecological. In terms of a computer analogy, life is both the data and the program and its biochemical structure is the way the information is embodied. This idea supports the seamless integration of life at all scales with the physical universe. The innovation reported here is essentially to integrate these ideas, basing information on the `general definition' of information, rather than simply the statistics of information, thereby explaining how functional information operates throughout life.

cs.IT↗

arXiv2012

Self-organizing traffic lights: A realistic simulation

We have previously shown in an abstract simulation (Gershenson, 2005) that self-organizing traffic lights can improve greatly traffic flow for any density. In this paper, we extend these results to a realistic setting, implementing self-organizing traffic lights in an advanced traffic simulator using real data from a Brussels avenue. On average, for different traffic densities, travel waiting times are reduced by 50% compared to the current green wave method.

nlin.AO↗

arXiv2012

Living in Living Cities

This paper presents an overview of current and potential applications of living technology to some urban problems. Living technology can be described as technology that exhibits the core features of living systems. These features can be useful to solve dynamic problems. In particular, urban problems concerning mobility, logistics, telecommunications, governance, safety, sustainability, and society and culture are presented, while solutions involving living technology are reviewed. A methodology for developing living technology is mentioned, while supraoptimal public transportation systems are used as a case study to illustrate the benefits of urban living technology. Finally, the usefulness of describing cities as living systems is discussed.

nlin.AO↗

arXiv2012

Complexity and Information: Measuring Emergence, Self-organization, and Homeostasis at Multiple Scales

Concepts used in the scientific study of complex systems have become so widespread that their use and abuse has led to ambiguity and confusion in their meaning. In this paper we use information theory to provide abstract and concise measures of complexity, emergence, self-organization, and homeostasis. The purpose is to clarify the meaning of these concepts with the aid of the proposed formal measures. In a simplified version of the measures (focusing on the information produced by a system), emergence becomes the opposite of self-organization, while complexity represents their balance. Homeostasis can be seen as a measure of the stability of the system. We use computational experiments on random Boolean networks and elementary cellular automata to illustrate our measures at multiple scales.

cs.IT↗

arXiv2012

Learning, Social Intelligence and the Turing Test - why an "out-of-the-box" Turing Machine will not pass the Turing Test

The Turing Test (TT) checks for human intelligence, rather than any putative general intelligence. It involves repeated interaction requiring learning in the form of adaption to the human conversation partner. It is a macro-level post-hoc test in contrast to the definition of a Turing Machine (TM), which is a prior micro-level definition. This raises the question of whether learning is just another computational process, i.e. can be implemented as a TM. Here we argue that learning or adaption is fundamentally different from computation, though it does involve processes that can be seen as computations. To illustrate this difference we compare (a) designing a TM and (b) learning a TM, defining them for the purpose of the argument. We show that there is a well-defined sequence of problems which are not effectively designable but are learnable, in the form of the bounded halting problem. Some characteristics of human intelligence are reviewed including it's: interactive nature, learning abilities, imitative tendencies, linguistic ability and context-dependency. A story that explains some of these is the Social Intelligence Hypothesis. If this is broadly correct, this points to the necessity of a considerable period of acculturation (social learning in context) if an artificial intelligence is to pass the TT. Whilst it is always possible to 'compile' the results of learning into a TM, this would not be a designed TM and would not be able to continually adapt (pass future TTs). We conclude three things, namely that: a purely "designed" TM will never pass the TT; that there is no such thing as a general intelligence since it necessary involves learning; and that learning/adaption and computation should be clearly distinguished.

cs.AI↗

arXiv2011

Facing Complexity: Prediction vs. Adaptation

One of the presuppositions of science since the times of Galileo, Newton, Laplace, and Descartes has been the predictability of the world. This idea has strongly influenced scientific and technological models. However, in recent decades, chaos and complexity have shown that not every phenomenon is predictable, even if it is deterministic. If a problem space is predictable, in theory we can find a solution via optimization. Nevertheless, if a problem space is not predictable, or it changes too fast, very probably optimization will offer obsolete solutions. This occurs often when the immediate solution affects the problem itself. An alternative is found in adaptation. An adaptive system will be able to find by itself new solutions for unforeseen situations.

nlin.AO↗

arXiv2011

Are Minds Computable?

This essay explores the limits of Turing machines concerning the modeling of minds and suggests alternatives to go beyond those limits.

cs.AI↗

arXiv2011

Complexity

The term complexity derives etymologically from the Latin plexus, which means interwoven. Intuitively, this implies that something complex is composed by elements that are difficult to separate. This difficulty arises from the relevant interactions that take place between components. This lack of separability is at odds with the classical scientific method - which has been used since the times of Galileo, Newton, Descartes, and Laplace - and has also influenced philosophy and engineering. In recent decades, the scientific study of complexity and complex systems has proposed a paradigm shift in science and philosophy, proposing novel methods that take into account relevant interactions.

cs.OH↗

arXiv2011

The Implications of Interactions for Science and Philosophy

Reductionism has dominated science and philosophy for centuries. Complexity has recently shown that interactions---which reductionism neglects---are relevant for understanding phenomena. When interactions are considered, reductionism becomes limited in several aspects. In this paper, I argue that interactions imply non-reductionism, non-materialism, non-predictability, non-Platonism, and non-nihilism. As alternatives to each of these, holism, informism, adaptation, contextuality, and meaningfulness are put forward, respectively. A worldview that includes interactions not only describes better our world, but can help to solve many open scientific, philosophical, and social problems caused by implications of reductionism.

nlin.AO↗

arXiv2011

Complex Networks

Introduction to the Special Issue on Complex Networks, Artificial Life journal.

cs.NE↗

arXiv2011

Polyethism in a colony of artificial ants

We explore self-organizing strategies for role assignment in a foraging task carried out by a colony of artificial agents. Our strategies are inspired by various mechanisms of division of labor (polyethism) observed in eusocial insects like ants, termites, or bees. Specifically we instantiate models of caste polyethism and age or temporal polyethism to evaluated the benefits to foraging in a dynamic environment. Our experiment is directly related to the exploration/exploitation trade of in machine learning.

cs.AI↗

arXiv2011

Self-organizing traffic lights at multiple-street intersections

Summary: Traffic light coordination is a complex problem. In this paper, we extend previous work on an abstract model of city traffic to allow for multiple street intersections. We test a self-organizing method in our model, showing that it is close to theoretical optima and superior to a traditional method of traffic light coordination. Abstract: The elementary cellular automaton following rule 184 can mimic particles flowing in one direction at a constant speed. This automaton can therefore model highway traffic. In a recent paper, we have incorporated intersections regulated by traffic lights to this model using exclusively elementary cellular automata. In such a paper, however, we only explored a rectangular grid. We now extend our model to more complex scenarios employing an hexagonal grid. This extension shows first that our model can readily incorporate multiple-way intersections and hence simulate complex scenarios. In addition, the current extension allows us to study and evaluate the behavior of two different kinds of traffic light controller for a grid of six-way streets allowing for either two or three street intersections: a traffic light that tries to adapt to the amount of traffic (which results in self-organizing traffic lights) and a system of synchronized traffic lights with coordinated rigid periods (sometimes called the "green wave" method). We observe a tradeoff between system capacity and topological complexity. The green wave method is unable to cope with the complexity of a higher-capacity scenario, while the self-organizing method is scalable, adapting to the complexity of a scenario and exploiting its maximum capacity. Additionally, in this paper we propose a benchmark, independent of methods and models, to measure the performance of a traffic light controller comparing it against a theoretical optimum.

nlin.AO↗

arXiv2011

Modular Random Boolean Networks

Random Boolean networks (RBNs) have been a popular model of genetic regulatory networks for more than four decades. However, most RBN studies have been made with random topologies, while real regulatory networks have been found to be modular. In this work, we extend classical RBNs to define modular RBNs. Statistical experiments and analytical results show that modularity has a strong effect on the properties of RBNs. In particular, modular RBNs have more attractors and are closer to criticality when chaotic dynamics would be expected, compared to classical RBNs.

nlin.CG↗

arXiv2010

The World as Evolving Information

This paper discusses the benefits of describing the world as information, especially in the study of the evolution of life and cognition. Traditional studies encounter problems because it is difficult to describe life and cognition in terms of matter and energy, since their laws are valid only at the physical scale. However, if matter and energy, as well as life and cognition, are described in terms of information, evolution can be described consistently as information becoming more complex. The paper presents eight tentative laws of information, valid at multiple scales, which are generalizations of Darwinian, cybernetic, thermodynamic, psychological, philosophical, and complexity principles. These are further used to discuss the notions of life, cognition and their evolution.

cs.IT↗

arXiv2010

Guiding the Self-organization of Random Boolean Networks

Random Boolean networks (RBNs) are models of genetic regulatory networks. It is useful to describe RBNs as self-organizing systems to study how changes in the nodes and connections affect the global network dynamics. This article reviews eight different methods for guiding the self-organization of RBNs. In particular, the article is focussed on guiding RBNs towards the critical dynamical regime, which is near the phase transition between the ordered and dynamical phases. The properties and advantages of the critical regime for life, computation, adaptability, evolvability, and robustness are reviewed. The guidance methods of RBNs can be used for engineering systems with the features of the critical regime, as well as for studying how natural selection evolved living systems, which are also critical.

nlin.AO↗

arXiv2010

Computing Networks: A General Framework to Contrast Neural and Swarm Cognitions

This paper presents the Computing Networks (CNs) framework. CNs are used to generalize neural and swarm architectures. Artificial neural networks, ant colony optimization, particle swarm optimization, and realistic biological models are used as examples of instantiations of CNs. The description of these architectures as CNs allows their comparison. Their differences and similarities allow the identification of properties that enable neural and swarm architectures to perform complex computations and exhibit complex cognitive abilities. In this context, the most relevant characteristics of CNs are the existence multiple dynamical and functional scales. The relationship between multiple dynamical and functional scales with adaptation, cognition (of brains and swarms) and computation is discussed.

cs.NE↗